h.264空間分辨率轉碼研究和實現(xiàn)

1、湖南大學碩士學位論文H.264空間分辨率轉碼研究和實現(xiàn)姓名：劉理申請學位級別：碩士專業(yè)：信號與信息處理指導教師：楊高波20060319H.264 分辨率轉碼研究和實現(xiàn)IIAbstractVideo transcoding is the operation of converting a compressed video stream to theother stream of different form for adapting to

2、 the network and usage environment. Theforms include the coding syntax, code rate, frame rate and spatial resolution. With theextensive use of the network appliances including handheld computers, personaldigital assistan

3、ts (PDAs), MP4, and smart cellular phones, which varies significantlyin resources such as computing power and display capability. To provide videocontents to these appliances, video transcoding is necessary. Among them,

4、spatialtranscoding is one of the most important kinds of transcoding. H.264 is the latestvideo coding standard, so this thesis focus on the H.264 related spatial transcoding.This paper first introduces the conception and

5、 classification of transcoding, whichincludes the conversion of bit-rate, spatial resolution, frame-rate and the code syntax.The features, key techniques and the search hot-topics of every category are brieflyintroduced.

6、 The main contributions of this thesis are detailed as follows: First, thearchitectures for spatial transcoding are surveyed, and the existing problems and keytechniques for implementation are summarized. Second, a motio

7、n re-estimationalgorithm based on hybrid rectangle and hexagon templates is proposed. It utilizes themotion vectors, which are easily obtained from the input video stream, to derive thedirectional derivative and gratitud

8、e, and then predicts the initial motion vectors.Motion re-estimation algorithm selects different search pattern of rectangle andhexagon template to obtain the proper the motion vectors. Third, a local directionbased fast

9、 mode selection algorithm is proposed. For INTRA model selection, itutilizes the histogram of the local prediction direction and selects that of highprobability. For INTER mode selection, it is based on the gradient to r

10、educe thenumber of candidate modes. Experimental results demonstrate that it can reduce thecomplexity of mode decision in downscaling video transcoding while keeping thequality of reconstructed video and coding efficienc